The support vector machine (SVM) is the best theory applied to
small sample classification and regression problems. It is based on the
structural risk minimization principle [15], and the decision rules are
obtained by a small number of training samples that are suitable for
the independent test set, can improve the generalization ability of the
learning machine, and can effectively solve the over-fitting problems.
In conclusion, SVM would become an alternative method to the
perceptron or neural network.